The judgment itself [PDF] is a fascinating read; quite apart from the details of this case, parts of it could serve as a lecture on the separation of powers, and the relationship between national security and the rule of law. Some parts of it also serve to illustrate that law isn’t as apolitical as some might think.
The argument, roughly, runs thus:
The SFO Director cited a threat to national security as the reason for stopping the investigation. But the word ‘threat’ is ambiguous. The normal sense in this sort of context is that of a risk or danger to be faced, but in this case:
representatives of a foreign state had issued a specific threat as to the consequences which would flow from a refusal to halt the investigation. It is one thing to assess the risk of damage which might flow from continuing an investigation, quite another to submit to a threat designed to compel the investigator to call a halt. [para 57]
This threat, that the Saudis would stop cooperating with UK counterterrorism intelligence if the investigation went ahead, “was not simply directed at this country’s commercial, diplomatic and security interests; it was aimed at its legal system” . This made it the business of the courts:
The courts are entitled to exercise their own judgment as to how best they may protect the rule of law, even in cases where it is threatened from abroad. In the exercise of that judgment we are of the view that a resolute refusal to buckle to such a threat is the only way the law can resist. 
Surrender deprives the law of any power to resist for the future.  …
Certainly, for the future, those who wish to deliver a threat designed to interfere with our internal, domestic system of law, need to be told that they cannot achieve their objective. 
This seems very unambiguous and uncompromising; regrettably so, because the ruling goes on to note the potential for legitimate exceptions. It accepts “that there may be circumstances so extreme that the necessity to save lives compels a decision not to detain or to prosecute”  – although it would be for the courts to rule on the legality of such a decision.
No attempt is made to specify exactly how such rulings should be made in general:
It is unnecessary for this court to attempt to identify those circumstances in which necessity may justify submission to a threat, designed to prevent a prosecutor from exercising his power to continue an investigation. 
But the ruling notes that:
there was no specific, direct threat made against the life of anyone. The threat made was to withdraw co-operation in relation to counter-terrorism. In order to assess the risk to life, it is necessary to hypothesise that a terrorist outrage was planned within the United Kingdom or elsewhere against British citizens or servicemen, of which Saudi Arabian intelligence had become aware and which it deliberately withheld. We readily accept that in 2006 and even now there is a serious risk of unpredictable terrorist attack, the greater the sources of intelligence the better that may be avoided and, as we are told, Saudi Arabia remains an irreplaceable source. But those factors do not, in our view, require us to accept that the Director was faced with [an adequate] degree of compulsion… 
And there is another relevant issue: “the courts are bound to question whether all the steps which could reasonably be taken to divert the threat had been pursued” . They see “no evidence whatever that any consideration was given as to how to persuade the Saudis to withdraw the threat, let alone any attempt made to resist the threat” .
Absence of evidence is not evidence of absence, though, so there may be protests that this is a case of suggesting the authorities be presumed guilty of negligence until proven innocent. On this, the judgment states:
The defendant and Government were well aware that the accusation was that they had surrendered too readily; it was for them to show not only that the consequences of the threat were dire but that the threat itself could not be mitigated or withdrawn. 
But even if this isn’t wholly convincing (and here I interpret for myself), this part of the argument isn’t vital. It is one of the two grounds on which the judgment rejects the SFO Director’s decision. If this one failed, the other (insufficient threat to national security) would still hold up.
Indeed, the judgment also contends that there is no “true distinction between national security and the rule of law”, as “preserving the rule of law constitutes an important component in the means by which democracy is secured” . Thus, in this case, “adequate consideration [was not] given to the damage to national security and to the rule of law by submission to the threat” . So, decisions such as these are not weighing national security against the rule of law but rather different aspects of national security against one another.
I wonder. Absent a clear definition of what constitutes sufficient danger to tolerate a subversion of the justice system, is there a conflict of interest in allowing members of that system to decide whether that subversion would be worse than other, non-judicial, dangers? There is always the danger that institutions with power will tend to gradually accrue more and to prioritise their sphere of influence over others.
For instance, there’s a notable jump between “Threats to the administration of public justice within the United Kingdom are the concern primarily of the courts, not the executive”  and “So too must the courts patrol the boundary between the territory which they safeguard and that for which the executive is responsible” . The latter seems to hint at the right not just to protect their territory but also to decide what belongs there.
(I’m reminded of Lord Bingham’s somewhat expansionary take on the definition of ‘the rule of law’.)
Just a thought. I’m also well aware of the opposing pressures, whereby politicians seek to restrict what the judiciary can do.
Anyway, the judgment is: threats intended to disrupt the rule of law must be resisted; there may be exceptions to this rule in awful situations; this case was not such a situation; furthermore, no evidence suggested that efforts had been made to divert the threat. The decision was an unlawful capitulation.
I think that’s fairly sound. All the same, I’m concerned about one principle that has now been written into legal chapter and verse, in paragraph 85 (quoted above). It noted that there was no “specific, direct threat”, and argued that “to assess the risk to life, it is necessary to hypothesise that a terrorist outrage was planned… of which Saudi Arabian intelligence had become aware”.
This seems to imply that any non-quantifiable danger, where specifics are not known, must be disregarded legally. Notions such as “a serious risk of unpredictable terrorist attack” and factors that will increase or reduce such a risk should not be written off so lightly.